Difference between revisions of "Private:vikas kumar"
From NMSL
(Submitted by Vikas on 26th June) |
|||
| Line 30: | Line 30: | ||
* We also want to support the longest IP-prefix matching if receiver and sender are in the same AS | * We also want to support the longest IP-prefix matching if receiver and sender are in the same AS | ||
* In addition to AS hop distance, also use IP hop distance to evaluate our ISP friendly matching. (Consider a couple of ISPs' graph at Rocketfuel project) | * In addition to AS hop distance, also use IP hop distance to evaluate our ISP friendly matching. (Consider a couple of ISPs' graph at Rocketfuel project) | ||
| + | |||
| + | == Week 4 + 5 Jun 15 == | ||
| + | * Completed AS Dynamics for core ASes( excluding customer only ASes ), for only providers ASes and also for others ASes( excluding only provider and only-customer ASes ). | ||
| + | * Implemented simulator for Random Matching, Network Order Matching and ISP Friendly Matching. | ||
| + | * Got the result for different population and different percentage popularity. We used real IPs for simulation of actual bit-torrent users. | ||
Revision as of 11:32, 26 June 2008
Work Progress for Vikas.
Priorities
- Implement Mao's Algorithm
- Implement our customer only heuristic
- Get familiar with Nitin's peer matching simulation
- Collect real IP addresses from Bit - Torrent networks for evaluation
- Quantify AS dynamics
- Validate valley-free rule
- more efficient peer matching algorithm
- Better shortest path algorithm
Week 1: May 26
- Rewriting the code for computing distance matrices between any pair of ASes based on Mao's Algorithm
Week 2 : Jun 2
- Completed implementation of Mao's Algorithm
- Completed implementation of customer only heuristic and verified that output of both implementations are the same
- Starting peer matching simulation using Nitin's simulator
Week 3: Jun 9
- Completed AS Dynamics
- Working on Nitin's Simulator
13th June's Meeting Minutes
- We want to compute AS Dynamics for only core ASes ( excluding customer only ASes)
- Implement the simulator for ISP friendly matching
- We also want to support the longest IP-prefix matching if receiver and sender are in the same AS
- In addition to AS hop distance, also use IP hop distance to evaluate our ISP friendly matching. (Consider a couple of ISPs' graph at Rocketfuel project)
Week 4 + 5 Jun 15
- Completed AS Dynamics for core ASes( excluding customer only ASes ), for only providers ASes and also for others ASes( excluding only provider and only-customer ASes ).
- Implemented simulator for Random Matching, Network Order Matching and ISP Friendly Matching.
- Got the result for different population and different percentage popularity. We used real IPs for simulation of actual bit-torrent users.
